Permeation, separation and reaction studies of palladium composite membranes
Palladium or palladium silver alloy composite membranes have been produced by electroless plating on either porous a-A!2 O 3 supports or porous stainless steel supports. It was found that the tin introduced into the membrane when using the tin-sensitisationactivation procedure decreased the durability of the membrane when being subjected to elevated temperatures. Two other activation procedures were used successfully for plating palladium on a-A!2 O3 supports. Membranes made via these new activation procedures could withstand more temperature cycles than those made via the tinsensitisation- activation procedure. The osmosis technique developed previously by other workers was used in this research to repair and produce membranes of higher quality. It was found that use of a very high osmotic pressure (3M NaCl aqueous solution) could improve the quality of a membrane by compressing the deposited palladium layer. Palladium-silver membranes were produced by plating followed by heattreatment, resulting in the formation of a palladium and silver alloy. The presence of hydrogen was helpful in this process. It was found experimentally that CO, H2 O, n-C4Hio and iso-C4Hio reduced the permeation of hydrogen through a palladium or palladium-silver composite membrane, while nitrogen had a neutral effect. Addition of CO and H2 O did not change any surface structure of the membrane, but retarded the permeation of hydrogen by competitive adsorption. Therefore, their effect disappeared after the membrane was purged by pure hydrogen for a short period of time. The adsorption tendency of H2 O was stronger than that of CO, so the former retarded the permeation of hydrogen more. The presence of n-C4Hio and iso-C4Hio caused the hydrogen permeability to decrease, and the permeability lost could not be completely recovered with a hydrogen purge. In the presence of dehydrogenation catalyst, the effect of these hydrocarbons became much more severe, and a hydrogen purge could not recover the permeability. This was due to the dehydrogenation product (olefin) causing some coke deposition on the membrane. Dilute oxygen/nitrogen mixtures could burn off the coke, leading to the recovery of the permeability. In the research, it was found that the hydrogen permeability of a composite palladium-silver membrane was affected by the permeation direction. For pure hydrogen, direction from the support to the metal film would give a higher hydrogen permeance, but for a hydrogen mixture with nitrogen, the result was just opposite. From the simulated results, it was found that the permeation direction from the metal film to the support could make the hydrogen permeation rate deviate from Sievert's law. The dehydrogenation of iso-butane was studied in a fixed bed and in a palladiumsilver membrane reactor. Pt/A^Os was used as the catalyst. While in a fixed bed reactor, total pressure reduced both the conversion of iso-butane and the selectivity of iso-butene, this was not the case in the palladium-silver membrane reactor. The removal of hydrogen from the reaction zone ensured that the experimental conversions of iso-butane in the membrane reactor were higher than the equilibrium values. The iso-butene selectivity was also increased in the membrane reactor and a flow rate of sweep gas had a favourable effect on the conversions and the selectivity. The hydrogen permeability of the membrane was much higher than the required hydrogen removal in this research, so the controlling step of the reaction was the activity of the catalyst. Counter-current sweep produced only slightly higher conversion than co-current sweep, but in the present work the co-current sweep mode was better than the counter-current sweep mode, because of a higher hydrogen partial pressure at the reactor end leading to less catalyst deactivation.